ARTIFICIAL INTELLIGENCE BASED SHOPPING TROLLEY FOR VISUALLY IMPAIRED PEOPLE WITH OPTICAL CHARACTER RECOGNITION

Abstract

Visually impaired individuals often face challenges in performing tasks such as shopping in supermarkets. To assist them, a proposed system utilizes camera-based text recognition and motion detection to identify objects and provide relevant information. The system employs Optical Character Recognition (OCR) to read text on product labels and AI-driven motion detection to sense the presence of nearby items. Audio instructions guide users, providing details about objects and their locations within the supermarket. This enables visually impaired shoppers to navigate the store independently and make informed choices. The system is designed to create a more convenient and accessible shopping experience, helping users move freely and comfortably in a sophisticated environment. By leveraging artificial intelligence and OCR technology, the solution simplifies shopping, saves time, and enhances autonomy for visually impaired individuals, making retail environments more inclusive and user-friendly.

Specification

Description:FIELD OF INVENTION
The user is interested in developing an artificial intelligence-based shopping trolley designed to assist visually impaired individuals. This trolley incorporates Optical Character Recognition (OCR) to identify product labels and prices, voice-guided navigation for in-store movement, and smart sensors to detect obstacles. The goal is to enhance shopping independence and accessibility, leveraging AI for real-time assistance and a seamless user experience.
BACKGROUND OF INVENTION
The Artificial Intelligence-Based Shopping Trolley for Visually Impaired People addresses the challenges faced by visually impaired individuals during shopping, offering them independence and convenience. Navigating a store, identifying products, reading labels, and managing a shopping list can be overwhelming for those with visual impairments. Traditional solutions, such as personal assistants or manual guidance, often compromise autonomy and privacy.
This invention leverages Artificial Intelligence (AI) and Optical Character Recognition (OCR) technologies to create a smart trolley that acts as a personal shopping assistant. OCR technology allows the trolley to read and interpret text on product labels and price tags, converting them into audio descriptions. Advanced AI algorithms process this information and provide real-time voice feedback, guiding users to make informed choices.
The trolley is equipped with sensors for obstacle detection, ensuring safe navigation through crowded aisles. A voice-controlled interface allows users to interact with the system, while preloaded shopping lists or dynamically updated instructions enhance efficiency. Integrated RFID or barcode scanning capabilities enable seamless identification of products for billing.
This innovation prioritizes user comfort and independence by blending cutting-edge AI, OCR, and IoT technologies. It eliminates the need for constant assistance, empowering visually impaired individuals to shop confidently and independently. Designed for scalability and affordability, the trolley can transform retail environments by promoting inclusivity and accessibility, enhancing the shopping experience for individuals with disabilities while fostering social equity.
The patent application number 201941023578 discloses a ergonomically designed multi-purpose trolley.
The patent application number 202021031186 discloses a radio frequency identification based shopping trolley.
The patent application number 202141001526 discloses a shopping trolley with automatic billing system.
The patent application number 202241042730 discloses a wearable device to assist blind and visually impaired people on indoor and outdoor environments.
The patent application number 202341006227 discloses a method and device for guiding a visually impaired user towards a target location.
SUMMARY
The Artificial Intelligence-Based Shopping Trolley for Visually Impaired People is an innovative solution that enhances the shopping experience for individuals with visual impairments, promoting independence, safety, and convenience. The invention integrates Artificial Intelligence (AI), Optical Character Recognition (OCR), and smart navigation technologies to assist users in identifying products, navigating stores, and making informed purchasing decisions.
The trolley is equipped with an OCR system that scans product labels and price tags, converting text into audio descriptions that are relayed to the user via a voice-guided interface. This enables visually impaired shoppers to identify products and access detailed information such as ingredients, pricing, and promotions. AI algorithms analyze this data in real-time, ensuring accurate and responsive feedback.
A key feature is the trolley's navigation system, powered by proximity sensors and object-detection technology, which identifies and avoids obstacles, enabling safe and smooth movement through crowded aisles. Voice commands allow users to interact with the trolley, request specific items, and receive directions to desired product locations. Additionally, the trolley can integrate with preloaded or cloud-synced shopping lists for personalized assistance.
Further enhancements include RFID or barcode scanning for product identification and billing, reducing manual effort at checkout. Designed for scalability and cost-effectiveness, the trolley aims to transform retail environments into inclusive spaces, ensuring accessibility for visually impaired individuals. This invention leverages technology to empower users, fostering independence and equality while redefining the shopping experience with intelligent, user-centric design.
DETAILED DESCRIPTION OF INVENTION
This system empowers visually impaired individuals to navigate stores independently without requiring assistance. It achieves this by reading product details using advanced text recognition techniques. The system boasts a 90% success rate on test datasets, depending on the size of the product text and the camera's distance. The proposed method involves capturing text, converting it into machine-readable format, and generating a text file. This file is processed using digital image analysis, which converts the extracted text into audio output. The audio instructions provide real-time information to users, enhancing their shopping experience. The system's primary goal is to improve the convenience of visually impaired individuals by delivering information through speech signals. A GPIO header is utilized in the design, mounted on the system board. Any GPIO pin can function as an input or output pin, ensuring versatile connectivity and robust system performance.

This system aids visually impaired individuals in identifying obstacle-free paths and provides essential information about their surroundings and products. Product details are conveyed to the user through vibrations, offering a tactile method of communication. The system's performance was evaluated in controlled settings and real-world environments using NavGuide, showing improved results compared to previous systems.
The NAVI system, described in this study, employs an RGB-D sensor to guide users via audio signals. This consumer-grade RGB-D camera integrates depth information and visual data, relying on factors such as image intensity and depth to navigate effectively.
Another approach involves electronic travel aids to detect obstacles and deliver surrounding information. Users receive real-time obstacle data through a haptic device. This prototype includes a short cane with an active handle and integrated sensing technology for precise guidance.
In a related study, a trolley system designed for visually impaired individuals simplifies movement within supermarkets. This trolley utilizes RFID technology, which surpasses the capabilities of barcode systems. RFID tags can be read from a distance of up to 300 feet, compared to only 15 feet with barcodes. Furthermore, RFID technology allows the system to process up to 40 tags simultaneously, enhancing efficiency and accuracy during shopping. These advancements highlight the potential of smart assistive systems in creating a more accessible and convenient experience for visually impaired users.

Figure 1: Wearable navigation
Wearable Navigation Systems for Visually Impaired Individuals
Wearable navigation systems are developed to assist visually impaired individuals in navigating various environments. These systems are categorized into three main types:
1. Electronic Travel Aids (ETAs): Devices that enhance users' perception of their surroundings.
2. Electronic Orientation Aids (EOAs): Tools that provide spatial and directional information.
3. Position Locator Devices (PLDs): Systems that determine the user's precise location.
These technologies focus on obstacle detection and avoidance, ensuring safer mobility for the users.
Electronic Mobility Cane (EMC)
The Electronic Mobility Cane (EMC) is another significant innovation that helps visually impaired users navigate by detecting obstacles, including staircases. It constructs a logical layout of the surrounding environment and uses intuitive vibration feedback to convey the proximity and priority of obstacles.
Image and Text Recognition for Audio Feedback
A system employing Raspberry Pi enhances the ability to convert images or text into machine-readable text and subsequently into speech. This technology assists visually impaired individuals in understanding their surroundings through audio descriptions.

Figure 2: Trolley movement and text recognition
System Architecture
1. Trolley Movement
The trolley's movement is powered by an L293D driver, connected to a Raspberry Pi. The L293D acts as a current amplifier, driving the trolley based on the input logic:
• 00 or 11: Trolley stops.
• 01 or 10: Trolley moves forward or backward.
2. Text Recognition
The system employs Optical Character Recognition (OCR) to scan and convert text from images into machine-readable text, which is then converted into speech for audio output.
Evaluation of the System
Advantages:
• The L293D driver facilitates movement in all directions.
• The system is cost-effective and extracts text from objects efficiently.
Disadvantages:
• The L293D driver chip has a voltage drop of 1.5V or more.
• OCR struggles with recognizing handwritten documents and text in low-light images.
Applications:
This technology can be implemented in various fields, including:
• Shopping: Assisting visually impaired users in identifying products.
• Banking: Text extraction for financial documentation.
• Healthcare and Government Agencies: Simplifying text processing and accessibility without manual assistance.

Figure 3: Flowchart
The Artificial Intelligence (AI)-based shopping trolley with Optical Character Recognition (OCR) is an innovative solution designed to assist visually impaired individuals in independently navigating and shopping in retail environments. By integrating advanced AI algorithms, OCR technology, and user-friendly features, the system empowers users by identifying products, providing real-time feedback, and ensuring accessibility.
The project highlights the potential of combining AI and assistive technologies to enhance the quality of life for visually impaired individuals. The system successfully addresses challenges such as product identification, mobility assistance, and accessibility in shopping spaces, enabling users to shop with greater confidence and autonomy.
This technology not only improves the shopping experience for visually impaired individuals but also serves as a foundation for future developments in assistive technologies, ensuring inclusivity and ease of use in daily life activities. Further refinement and expansion of such solutions can pave the way for broader applications across various domains, fostering a more inclusive society.

DETAILED DESCRIPTION OF DIAGRAM
Figure 1: Wearable navigation
Figure 2: Trolley movement and text recognition
Figure 3: Flowchart , Claims:1. Artificial Intelligence Based Shopping Trolley for Visually Impaired People with Optical Character Recognition claims that the AI-based trolley enables visually impaired individuals to independently navigate and shop, removing the need for constant assistance.
2. Integrated Optical Character Recognition (OCR) accurately identifies product names, descriptions, and pricing, making the shopping process seamless.
3. The trolley uses real-time voice feedback to communicate product details, guiding users effectively during their shopping experience.
4. Advanced AI algorithms ensure obstacle detection and safe navigation within the store, reducing the risk of accidents.
5. The system supports user customization, allowing adjustments for language preferences and personalized shopping needs.
6. Provides instant feedback on scanned items, enhancing user confidence and efficiency in selecting desired products.
7. The trolley is designed for ease of use, with ergonomic considerations to accommodate various users and store layouts.
8. The solution can be integrated into a wide range of retail environments, supporting diverse inventories and layouts.
9. By using commercially available components and scalable AI technology, the system offers an affordable assistive device for visually impaired individuals.
10. By empowering visually impaired shoppers, the trolley fosters independence and inclusivity, contributing to a more equitable shopping environment.